J Am Soc Nephrol 14: S65–S70, 2003

Risk Factors for Progressive Chronic Kidney Disease WILLIAM M. MCCLELLAN and W. DANA FLANDERS Department of Epidemiology, Rollins School of Public Health, Emory University, Atlanta, Georgia.

Abstract. The occurrence of chronic kidney disease and subsequent rate of loss of renal function are highly variable among individuals with the same underlying cause of renal injury or degree of functional impairment. Individual variability of risk is typical of complex diseases and reflects the multifactorial nature of the biologic mechanisms that are involved in the underlying disease process. The utility of the risk factor concept in developing CKD prevention and control strategies includes identifying individuals at high risk for the occurrence and progression of CKD, defining at-risk populations, elucidating potential targets for intervention, and generating explan-

atory hypotheses for the variable risk of CKD noted in different populations. Future application of the risk factor concept in the prevention and control of CKD will entail developing multivariate prediction equations; using spatial and temporal, as well as personal, characteristics, to define at-risk populations; identifying biomarkers for complex risk factors like race; and translating this information into testable interventions. This should include active extension of our current understanding of health care, social, and economic risk factors at both the individual and the community level.

The incidence and progression of renal injury vary substantially among individuals who are at-risk for kidney disease. For example, 8% of new patients with type 2 diabetes mellitus already have proteinuria at diagnosis (1). Among patients with type 2 diabetes mellitus who are initially free of proteinuria, the 20-yr risk of diabetic nephropathy is 41% (1). After proteinuria occurs, the subsequent 10-yr risk of progressive chronic kidney disease is 11% (2). Thus, about half of those with type 2 diabetes will develop nephropathy and 10% of these individuals will experience progressive loss of renal function. Variable risk of impaired renal function has also been reported among hypertensive subjects. At study entry, 5.9% of the Hypertension Detection and Follow-up Program trial participants had a serum creatinine of 1.5 mg/dl or greater and 2.3% of the 8683 participants with serial serum creatinine measurements over 5 yr experienced clinically significant loss of renal function (3). Similarly, the rate of loss of renal function among individuals with CKD also displays considerable person-to-person variability. For example, the Modification of Diet in Renal Disease (MDRD) study reported that the mean (SD) rate of decline in GFR during follow-up was 3.8 (⫾ 4.2) ml/min per yr for patients with moderate and 4.0 (⫾ 3.1) ml/min per yr for those with severe renal insufficiency at baseline (4). Substantial proportions of each group, 19% and 11%, respectively, showed no progression during follow-up, further indicating variability in progression.

Risk Factors and Variable Risk of CKD

Correspondence to Dr. William McClellan, Georgia Medical Care Foundation, 57 Executive Park South, NE, Suite 200, Atlanta, GA. Phone: 404-982-7573; Fax: 678-527-3473 E-mail: [email protected] 1046-6673/1407-0065 Journal of the American Society of Nephrology Copyright © 2003 by the American Society of Nephrology DOI: 10.1097/01.ASN.0000070147.10399.9E

Risk Factors and Variable Risk. Variability of risk for the occurrence and progression of CKD suggests that biologically relevant characteristics exist that influence the occurrence or course of the renal disease. The variable progression of CKD among individuals with autosomal dominant polycystic kidney disease (ADPKD) illustrates this concept. The mutation of the polycystin gene that occurs within family members is a necessary risk factor for ADPKD. However, family members who inherit the same mutation demonstrate highly variable rates of progression to ESRD (5). The within-family heterogeneity of time of onset of ESRD indicates that the rate of progression of CKD among individuals with the same ADPKD gene is associated with one or more independent exposures other than the mutation itself. Combinations of causal factors that result in rapid progression to ESRD define sufficient causes for early loss of renal function among ADPKD patients (6). Although some or all of the other component causes for rapid progression of ADPKD may also be genetic, this formulation of necessary and sufficient causes tempers inclinations to monogenic determinism and emphasizes the multifactorial nature of most, if not all, diseases. The risk factor concept emphasizes that an exposure may be causal or may simply be associated with some other, perhaps unidentified, causal factor.1 Evidence supporting a potential causal role for a risk factor includes strong, graded associations with the disease that are consistent across different populations and study designs. A biologic link between the risk factor and the disease, relevant animal models, and strong dose-response relationships can also provide support of a causal role for a risk

1

The definition of risk factor is inconsistent within the epidemiologic literature and some authors reserve the term for causal factors (8). Further, risk prediction equations like those from the Framingham Heart Study for cardiovascular disease risk, typically use both well defined causal factors like serum cholesterol levels and poorly specified risk factors like gender.

S66

Journal of the American Society of Nephrology

factor. However, in many instances neither the presence of association nor the presence of strong supporting evidence is sufficient to establish causality and, often, causal evidence must rest on outcomes from randomized clinical trials. Even though a risk factor need not be causal, the concept has considerable utility in developing disease prevention and control strategies. First, risk factors can be used to identify individuals at high risk for the occurrence and progression of CKD. Second, risk factors can identify at-risk populations. Third, risk factors can identify potential new targets for intervention. Fourth, risk factors generate explanatory hypotheses. Each of these applications of the risk factor model to CKD warrants further comment.

Prediction of Risk Prediction of increased risk of occurrence or progression of CKD enables clinicians to identify individuals who may benefit from closer supervision of care or more intensive disease modifying interventions. Risk stratification using single patient characteristics is a straightforward application of this concept. Extending this concept to multiple risk factors requires statistical models that combine the contribution of multiple factors into a single summary score. For chronic kidney disease, interest has focused on risk factors for its occurrence and rate of progression. Information about these risk factors can be used in risk prediction (7,8).

Risk for Occurrence Associations between patient characteristics and the occurrence of kidney disease identify risk factors for disease occurrence. For a factor that increases risk, the probability of disease when the factor is present exceeds that in absence of the characteristic

p (disease 兩 characteristic present) ⬎ p (disease 兩 characteristic absent) Multivariate logistic regression models can account for the joint effects of multiple factors on the occurrence of CVD (19,20). The multivariate logistic model provides an estimation of risk for subsequent disease:

p (disease 兩 risk factors) ⫽ exp{␣ ⫹ ¥i[␤i (risk factori)]}/ 1 ⫹ exp{␣ ⫹ ¥i关␤i共risk factori)]} where p is the probability of CKD during a stipulated period of observation, ␣ and ␤i are logistic regression parameters derived from data for the population of interest, risk factor i is the individual’s value for the ith characteristic and the summation is over risk factors i ⫽ 1, 2, 3, . . . n. Although a widely accepted risk prediction model for the occurrence of CKD has not yet been developed for CKD, a recent cohort study illustrates a type of study design from which a prediction model could be derived (9). The study examined the 10-yr risk of developing ESRD among 107,192 participants aged 18 yr and older at baseline. ESRD occurred in 193 patients (0.18%). Patient characteristics considered as can-

J Am Soc Nephrol 14: S65–S70, 2003

didate risk factors for progressive CKD included gender, age at screening, proteinuria, hematuria, and systolic and diastolic BP. An independent association (adjusted odds ratio, 95% confidence interval) was found between risk of ESRD and proteinuria (14.9, 10.9 to 20.2), hematuria (2.30, 1.62 to 3.28), male gender (1.41, 1.04 to 1.92), and diastolic BP (1.39, 1.17 to 1.64).

Risk for Rate of Progression Information about the rate of change in renal function over time can serve as a measure of disease progression. Progression rates, which measure change in renal function per unit of time, should be clearly distinguished from incidence rates for CKD, which are measured in incident cases per person-year. Progression of CKD can be measured using serum creatinine (10), reciprocal of the serum creatinine (11), Cockcroft-Gault estimated GFR (or similar estimating equation) (12), creatinine clearance (13), or log-normalized albumin excretion rate (14). At present there is no consensus as to which of these measures is best for risk prediction. Risk factors for progression can be defined in a manner similar to that for occurrence:

Rate of change 兩 risk factor present ⬎ Rate of change 兩 risk factor absent Multivariate linear models can extend the simple comparison of rates in two groups to account for multiple risk factors. Finally, individual rates of progression can be predicted by the multivariate model:

Predicted rate ⫽ ␣ ⫹ ¥i␤i (risk factori) where the predicted rate is the expected decline in renal function, ␣ and ␤i are linear regression parameters derived from the data for the population of interest, and risk factors i the individual’s values for the characteristics included in the model. Because renal function is measured serially for each patient, it is necessary to account for the correlated nature of the outcome data using a linear growth curve or similar models (15). A widely accepted prediction model for the rate of progression of CKD has not yet been developed. A report from the MDRD trial illustrates how an equation to predict progression of CKD might be derived (4). The MDRD investigators examined 41 potential risk factors for loss of renal function as measured by 125I-iothalamate GFR. Factors independently associated with the progression of CKD included black race, increased mean arterial BP, baseline urine protein excretion, a diagnosis of polycystic kidney disease, and lower baseline levels of serum transferrin and HDL cholesterol. These six factors accounted for 34.5% of the variation in the rate of progression in mild and 33.9% in moderate renal insufficiency.

Potential Targets for Intervention The search for risk factors for CKD seeks to identify modifiable processes or mediators responsible for the occurrence and progression of renal failure. Table 1 lists some known risk

J Am Soc Nephrol 14: S65–S70, 2003

Risk Factors for Progressive Chronic Kidney Disease

S67

Table 1. Factors associated with the occurrence or progression of CKD

which can be tested in RCT. Examples of biomarkers associated with increased risk of CKD are listed in Table 1.

CKD Risk Factors older age (14,16) race and ethnicity (17–19) gender (20) lower birth weight (21) low socioeconomic status (22) smoking (23) alcohol consumption (24) familial aggregation (25) lead and other heavy metals (26) analgesic abuse (27) illicit drug use (28) dietary phytoestrogens (29) Biomarkers hemoglobin (30) oxidative stress/Carbonyl stress (31) Insulin Resistance Syndrome (32) hyperlipidemia (33) proteinuria (34) Genetic Markers Access to Adequate Health Care high blood pressure (35, 36) poorly controlled diabetes (37–39) ACE inhibitor use (40) access to health care (22) Social System poverty (41, 42)

At-Risk Populations

factors for renal failure. Some of these factors, like the use of angiotensin converting enzyme inhibitors, BP control and glycemic control, have been tested in randomized clinical trials (RCT) that have established them as disease-modifying interventions and suitable targets for intervention. Others, like analgesic abuse and smoking, are supported by strong observational evidence. Although RCT evidence is lacking that cessation of smoking or of chronic analgesic use is renoprotective, it is reasonable to suggest that interventions targeted at these risk factors, particularly in high-risk populations, can be supported. Other factors, like anemia, are currently the subjects of large, well designed RCT; recommendations for specific intervention await supportive evidence from the completion of these trials. In contrast, many of the risk factors listed in Table 1 like race, family membership, and socioeconomic status are complex measures and the biologic mechanisms that link them to the occurrence and progression of CKD are not defined. One challenge posed by these poorly characterized risk factors is to find associated biomarkers that directly link them to disease processes. Biomarkers are measured at the tissue, cellular, subcellular, molecular, or genetic level (43). They can potentially reduce misclassification and measurement variation. Furthermore, biomarkers may be more indicative of relevant disease processes that can suggest therapeutic interventions,

Risk factors can be used to define at-risk population that can be targeted for education and early intervention programs. Groups defined by race are often mentioned as at-risk populations. Blacks, for example, have a considerably higher risk of ESRD partially due to hypertension and diabetes. This increased risk peaks during early adulthood and cannot be fully accounted for by racial differences in the underlying prevalence of hypertension (44). A recent study by Kiberd and Clase estimated the cumulative risk of ESRD among different gender-race groups (45). They found that the lifetime of ESRD for a 20-yr-old black woman was 7.8%; for black males, 7.3%; for white women, 1.8%; and for white males, 2.5%. The number of years of life lost due to ESRD was comparable to the loss due to breast cancer among black women and prostate cancer among black men. Space and time can also define at-risk populations. For example, maps of race- and cause-specific ESRD incidence published by the United States Renal Data System reveal a substantial nonrandom distribution of high risk populations for diabetes and hypertension (46). Those geographic areas where higher ESRD rates are found might be considered for targeted interventions. These geographic variations in risk also suggest that there may be environmental and social determinants of CKD risk, which are not accounted for within the traditional risk factor model. A recent study by Lopes and Port of the age-specific patterns of black to white risk of nondiabetic ESRD illustrates the use of time as a risk factor to identify high-risk groups (47). They found that the relative risk of blacks compared to whites for CKD demonstrated considerable variability with age, peaking in middle adulthood and then declining, suggesting that blacks in midlife experience greater absolute and relative risk of CKD than at other times during their life span.

Explanatory Hypotheses Risk factors generate explanatory hypotheses that can lead to potential interventions. One set of explanatory hypotheses seeks to reduce complex exposures to potential causal factors. Well defined CKD risk factors that display a strong, graded association with the occurrence of CKD, which is consistent across studies in different populations and is related to biologic mechanisms responsible for renal injury and progressive loss of function, identify potential targets for disease modifying interventions. However, as noted above, many CKD risk factors are complex and not well defined. In such cases, it is necessary to consider how they might be linked to potential interventions. A second explanatory issue revolves around the amount of the risk within a population that can be accounted for by a set of risk factors, a measure called population attributable risk (48). Questions of this nature seek to quantify the proportion of risk that can be prevented by eliminating a set of known causal risk factors. Understanding the amount of disease within a

S68

Journal of the American Society of Nephrology

population attributable to one or more risk factors can facilitate a more appropriate allocation of resources for prevention programs.

Race as a Complex Risk Factor An approach to dealing with a complex exposure is to identify, separately measure, and study the components of the exposure. Even complex risk factors like race can be related directly or indirectly to genetic and molecular events responsible for the pathogenesis and progression of kidney disease. For example, the increased risk of ESRD associated with race has generated multiple explanatory hypotheses. Examples include studies of misclassification of hypertensive renal disease (49,50), racial differences in birth weight (51), familial aggregation of risk (25), variations in mitochondrial genes (52), poverty (22), and differential access to health care (22). Many of these risk factors are in themselves complex and subject to reduction component factors. The goal of reduction toward simpler component factors is to identify molecular biomarkers or environmental factors that account for the observed risk and that can be linked to biologically relevant disease mechanisms (43).

J Am Soc Nephrol 14: S65–S70, 2003

Population-Attributable Risk A recent report by Tarver-Carr et al. illustrates how multivariate risk models can be used to estimate the fraction of excess risk for CKD among African-Americans that can be attributed to risk factors other than race (59). They used data from the second National Health and Nutrition Examination survey (NHANES II) follow-up study to examine risk factors associated with higher risk of ESRD among African Americans compared with whites. They adjusted black to white relative risk estimates for: (1) age and gender alone; (2) sociodemographic factors; (3) lifestyle factors; and (4) clinical factors. They found that the proportion of the black versus white excess risk attributable to sociodemographic (poverty status, educational level, and marital status) factors was 12%; for lifestyle factors, 24%; and for clinical factors, 32%. All studied risk factors accounted for 44% of the excess risk for CKD experienced by African Americans. These results suggest that the factors included in the sociodemographic, lifestyle and clinical risk sets represent important intervention targets, and that substantial variation in black to white risk of ESRD remains to be explained.

Context as a Complex Risk Factor

Conclusion

Characteristics measured at the group level that are associated with increased risk are called contextual factors. The geographic variation in risk of ESRD described above is an example of a contextual risk factor. As one of many hypotheses, it has been suggested that economically and environmentally disadvantaged neighborhoods may account for geographic variations in risk (53,54). For example, Whittle et al. examined the association between ESRD incidence rates due to hypertension and neighborhood level measures of socioeconomic status derived from a random household survey (55). They found an ecological association—that is, one occurring at the group level—wherein communities with lower educational and income levels tended to have higher ESRD incidence rates. The inference of an association based on ecological observations is subject to ecologic bias if the exposure (low SES) and the CKD occur in different segments of the population (56). In a case-control study, the same group reported increased risk of ESRD associated with individual measures of socioeconomic status (household annual income and years of education) and with measures of health care access (health insurance status, dental health, usual source of health care, and use of preventive health care) (22). However, this study used individual level information and failed to capture the ecological information about risk that was observed in the population-based study. One way of dealing with this limitation is to include measures of SES and access to health care at both the individual and group level in a multilevel model. The statistical models should account for the complex correlated data structure that arises when individuals within communities are studied (57). A recent article by Diez Roux et al. examining neighborhood environments, individual socioeconomic status, and risk of coronary heart disease illustrates this approach to contextual, multilevel models of risk (58).

Increasing our understanding of risk factors for the occurrence and progression of CKD poses substantial challenges. Predictive models for risk and rates of progression based on this information should be derived and validated and their clinical applications studied. Biomarkers should be sought to clarify the associations between complex risk factors and risk of CKD. Once a risk factor has been fully characterized, therapeutic interventions based on this understanding should be devised and tested with randomized clinical trials. We should actively extend our understanding of health care and social and economic factors at the individual level and at the contextual level.

References 1. Larson TS, Santanello N, Shahinfar S, O’Brien PC, Palumbo PJ, Melton LJ 3rd, Leibson CL: Trends in persistent proteinuria in adult-onset diabetes: A population-based study. Diabetes Care 23: 51–56, 2000 2. Humphrey LL, Ballard DJ, Frohnert PP, Chu CP, O’Fallon WM, Palumbo PJ: Chronic renal failure in non-insulin-dependent diabetes mellitus. A population-based study in Rochester, Minnesota. Ann Intern Med 111: 788 –796, 1989 3. Shulman NB, Ford CE, Hall WD, Blaufox MD, Simon D, Langford HG, Schneider KA: Prognostic value of serum creatinine and effect of treatment of hypertension on renal function. Results from the hypertension detection and follow-up program. The Hypertension Detection and Follow-up Program Cooperative Group. Hypertension 13(Suppl I): 80 –93, 1989 4. Hunsicker LG, Adler S, Caggiula A, England BK, Greene T, Kusek JW, Rogers NL, Teschan PE: Predictors of the progression of renal disease in the Modification of Diet in Renal Disease Study. Kidney Int 51: 1908 –1919, 1997 5. Gabow PA, Johnson AM, Kaehny WD, Kimberling WJ, Lezotte DC, Duley IT, Jones RH: Factors affecting the progression of

J Am Soc Nephrol 14: S65–S70, 2003

6. 7.

8.

9.

10.

11.

12.

13.

14.

15. 16. 17. 18.

19.

20.

21.

22.

23. 24.

25.

renal disease in autosomal-dominant polycystic kidney disease. Kidney Int 41: 1311–1319, 1992 Rothman KJ: Causes. Am J Epidemiol 104: 587–592, 1976 Kleinbaum DG, Kupper LL, Morgenstern H: Epidemiologic Research: Principles and Quantitative Methods, Belmont, CA, Lifetime Learning Publications, 1982 Kannel WB, Larson M: Long-term epidemiologic prediction of coronary disease: The Framingham experience. Cardiology 82: 137–152, 1993 Iseki K, Iseki C, Ikemiya Y, Fukiyama K: Risk of developing end-stage renal disease in a cohort of mass screening. Kidney Int 49: 800 – 805, 1996 Perneger TV, Nieto FJ, Whelton PK, Klag MJ, Comstock GW, Szklo M: A prospective study of blood pressure and serum creatinine. Results from the “Clue” Study and the ARIC Study. JAMA 269: 488 – 493, 1993 Gabow PA, Johnson AM, Kaehny WD, Kimberling WJ, Lezotte DC, Duley IT, Jones RH: Factors affecting the progression of renal disease in autosomal-dominant polycystic kidney disease. Kidney Int 41: 1311–1319, 1992 Mazouz H, Kacso I, Ghazali A, El Esper N, Moriniere P, Makdassi R, Hardy P, Westeel PF, Achard JM, Pruna A, Fournier A: Risk factors of renal failure progression two years prior to dialysis. Clin Nephrol 51: 355–366, 1999 Jungers P, Hannedouche T, Itakura Y, Albouze G, DescampsLatscha B, Man NK: Progression rate to end-stage renal failure in non-diabetic kidney diseases: A multivariate analysis of determinant factors. Nephrol Dial Transplant 10: 1353–1360, 1995 Goetz FC, Jacobs DR Jr, Chavers B, Roel J, Yelle M, Sprafka JM: Risk factors for kidney damage in the adult population of Wadena, Minnesota: A prospective study. Am J Epidemiol 145: 91–102, 1997 Liang KY, Zeger SL: Longitudinal data analysis a generalized linear models. Biometrika 73: 13–22, 1986 Rodriguez-Puyol D: The aging kidney. Kidney Int 54: 2247– 2265, 1998 Rostand SG: US minority groups and end-stage renal disease: A disproportionate share. Am J Kidney Dis 19: 411– 413, 1992 Pugh JA, Stern MP, Haffner SM, Eifler CW, Zapata M: Excess incidence of treatment of end-stage renal disease in Mexican Americans. Am J Epidemiol 127: 135–144, 1988 Hoy WE, Megill DM: End-stage renal disease in southwestern Native Americans, with special focus on the Zuni and Navajo Indians. Transplant Proc 21: 3906 –3908, 1989 Neugarten J, Acharya A, Silbiger SR: Effect of gender on the progression of nondiabetic renal disease: A meta-analysis. J Am Soc Nephrol 11: 319 –329, 2000 Lackland DT, Bendall HE, Osmond C, Egan BM, Barker DJ: Low birth weights contribute to high rates of early-onset chronic renal failure in the Southeastern United States. Arch Intern Med 160: 1472–1476, 2000 Perneger TV, Whelton PK, Klag MJ: Race and end-stage renal disease. Socioeconomic status and access to health care as mediating factors. Arch Intern Med 155: 1201–1208, 1995 Orth SR, Ritz E, Schrier RW: The renal risks of smoking. Kidney Int 51: 1669 –1677, 1997 Perneger TV, Whelton PK, Puddey IB, Klag MJ: Risk of endstage renal disease associated with alcohol consumption. Am J Epidemiol 150: 1275–1281, 1999 Freedman BI, Soucie JM, McClellan WM: Family history of end-stage renal disease among incident dialysis patients. J Am Soc Nephrol 8: 1942–1945, 1997

Risk Factors for Progressive Chronic Kidney Disease

S69

26. Lin JL, Tan DT, Hsu KH, Yu CC: Environmental lead exposure and progressive renal insufficiency. Arch Intern Med 161: 264 – 271, 2001 27. McLaughlin JK, Lipworth L, Chow WH, Blot WJ: Analgesic use and chronic renal failure: A critical review of the epidemiologic literature. Kidney Int 54: 679 – 686, 1998 28. Norris KC, Thornhill-Joynes M, Robinson C, Strickland T, Alperson BL, Witana SC, Ward HJ: Cocaine use, hypertension, and end-stage renal disease. Am J Kidney Dis 38: 523–528, 2001. 29. Velasquez MT, Bhathena SJ: Dietary phytoestrogens: A possible role in renal disease protection. Am J Kidney Dis 37: 1056 – 68, 2001 30. Rossert JA, McClellan WM, Roger SD, Verbeelen DL, Horl WH: Contribution of anemia to progression of renal disease: A debate. Nephrol Dial Transplant 17[Suppl 1]: 60 – 66, 2002 31. Haugen E, Nath KA: The involvement of oxidative stress in the progression of renal injury. Blood Purif 17: 58 – 65, 1999 32. Manitius J, Biedunkiewicz B, Kustosz J, Rutkowski B: The relationship between insulin, glucose and serum uric acid and their contribution to the progression of renal damage in patients with primary glomerulonephritis. J Int Med Res 24: 449 – 453, 1996 33. Fried LF, Orchard TJ, Kasiske BL: Effect of lipid reduction on the progression of renal disease: A meta-analysis. Kidney Int 59: 260 –269, 2001 34. Jafar TH, Stark PC, Schmid CH, Landa M, Maschio G, Marcantoni C, de Jong PE, de Zeeuw D, Shahinfar S, Ruggenenti P, Remuzzi G, Levey AS: Proteinuria as a modifiable risk factor for the progression of non-diabetic renal disease. Kidney Int 60: 1131–1140, 2001 35. Luke RG: Hypertensive nephrosclerosis: pathogenesis and prevalence. Essential hypertension is an important cause of end-stage renal disease. Nephrol Dial Transplant 14: 2271–2278, 1999 36. Zucchelli P, Zuccala A: The kidney as a victim of essential hypertension. J Nephrol 10: 203–206, 1997 37. Anonymous: Retinopathy and nephropathy in patients with type 1 diabetes four years after a trial of intensive therapy. The Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications Research Group. N Engl J Med 342: 381–389, 2000 38. Stratton IM, Adler AI, Neil HA, Matthews DR, Manley SE, Cull CA, Hadden D, Turner RC, Holman RR: Association of glycaemia with macrovascular and microvascular complications of type 2 diabetes (UKPDS 35): Prospective observational study. Br Med J 321: 405– 412, 2000 39. UK Prospective Diabetes Study (UKPDS) Group: Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33). Lancet 352: 837– 853, 1998 40. Kasiske BL, Kalil RS, Ma JZ, Liao M, Keane WF: Effect of antihypertensive therapy on the kidney in patients with diabetes: a meta-regression analysis. Ann Intern Med 118: 129 –138, 1993 41. Young EW, Mauger EA, Jiang KH, Port FK, Wolfe RA: Socioeconomic status and end-stage renal disease in the United States. Kidney Int 45: 907–911, 1994 42. Byrne C, Nedelman J, Luke RG: Race, socioeconomic status, and the development of end-stage renal disease. Am J Kidney Dis 23: 16 –22, 1994 43. Schulte PA: A conceptual and historical framework for molecular epidemiology. In: Molecular Epidemiology: Principles and Practice, edited by Schulte PA and Perera FP, New York, Academic Press, 1993

S70

Journal of the American Society of Nephrology

44. McClellan W, Tuttle E, Issa A: Racial differences in the incidence of hypertensive end-stage renal disease (ESRD) are not entirely explained by differences in the prevalence of hypertension. Am J Kidney Dis 12: 285–290, 1988 45. Kiberd BA, Clase CM: Cumulative risk for developing end-stage renal disease in the US population. J Am Soc Nephrol 13: 1635–1644, 2002 46. United States Renal Data System: 2001 Annual Data Report: Atlas of end-stage renal disease in the United States, Bethesda, MD, National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases, 2001 47. Lopes AA, Hornbuckle K, James SA, Port FK: The joint effects of race and age on the risk of end-stage renal disease attributed to hypertension. Am J Kidney Dis 24: 554 –560, 1994 48. Rothman KJ: Modern Epidemiology, Boston, Little, Brown and Company, 1986 49. Perneger TV, Whelton PK, Klag MJ, Rossiter KA: Diagnosis of hypertensive end-stage renal disease: Effect of patient’s race. Am J Epidemiol 141: 10 –15, 1995 50. Blythe WB, Maddux FW: Hypertension as a causative diagnosis of patients entering end-stage renal disease programs in the United States from 1980 to 1986. Am J Kidney Dis 18: 33–37, 1991 51. Lopes AA, Port FK: The low birth weight hypothesis as a plausible explanation for the black/white differences in hypertension, non-insulin-dependent diabetes, and end-stage renal disease. Am J Kidney Dis 25: 350 –356, 1995

J Am Soc Nephrol 14: S65–S70, 2003

52. Watson B Jr, Khan MA, Desmond RA, Bergman S: Mitochondrial DNA mutations in black Americans with hypertensionassociated end-stage renal disease. Am J Kidney Dis 38: 529 – 536, 2001 53. Diez-Roux AV: Bringing context back into epidemiology: Variables and fallacies in multilevel analysis. Am J Public Health 88: 216 –222, 1998 54. Yen IH, Syme SL: The social environment and health: A discussion of the epidemiologic literature. Annu Rev Public Health 20: 287–308, 1999 55. Whittle JC, Whelton PK, Seidler AJ, Klag MJ: Does racial variation in risk factors explain black-white differences in the incidence of hypertensive end-stage renal disease? Arch Intern Med 151: 1359 –1364, 1991 56. Greenland S, Morgenstern H: Ecological bias, confounding, and effect modification. Int J Epidemiol 18: 269 –274, 1989 57. Diez Roux AV: Investigating neighborhood and area effects on health. Am J Public Health 91: 1783–1789, 2001 58. Diez Roux AV, Merkin SS, Arnett D, Chambless L, Massing M, Nieto FJ, Sorlie P, Szklo M, Tyroler HA, Watson RL: Neighborhood of residence and incidence of coronary heart disease. N Engl J Med 345: 99 –106, 2001 59. Tarver-Carr ME, Powe NR, Eberhardt MS, LaVeist TA, Kington RS, Coresh J, Brancati FL: Excess risk of chronic kidney disease in African-Americans versus whites in the United States: A population-based study of potential explanatory factors. JASN, in press